Technical Field
The present disclosure relates to a process for manufacturing MEMS (Micro-Electro-Mechanical System) devices having two buried cavities and to the micro-electro-mechanical device.
Description of the Related Art
As is known, sensors including micromechanical structures made, at least in part, of semiconductor materials employing MEMS technology are increasingly widely used, due to their advantageous characteristics of small dimensions, low manufacturing costs, and flexibility.
A MEMS sensor generally comprises a micromechanical sensing structure, which transduces a physical or mechanical quantity to be detected into an electrical quantity (for example, correlated to a capacitive variation); and an electronic reading circuit, usually formed as an ASIC (Application-Specific Integrated Circuit), which carries out processing operations (i.e., amplification and filtering) of the electrical quantity and supplies an electrical output signal, either analog (for example, a voltage) or digital (for example a PDM—Pulse Density Modulation—signal). The electrical signal, possibly further processed by an electronic interface circuit, is then made available to an external electronic system, for example a microprocessor control circuit of an electronic apparatus incorporating the sensor.
MEMS sensors comprise, for example, sensors for detecting physical quantities, such as inertial sensors, which detect acceleration or angular velocity data; sensors of derived signals, such as quaternions (data representing rotations and directions in three-dimensional space), gravity signals, etc.; motion sensors, such as step counters, running sensors, uphill sensors, etc.; and environmental signals, which detect quantities such as pressure, temperature, and humidity.
To sense the physical/mechanical quantity, MEMS sensors of the considered type comprise a membrane or a mass formed in or on a semiconductor chip and suspended over a first cavity. The membrane may face the external environment or be in communication therewith via a fluidic path.
U.S. Pat. No. 9,233,834 describes, for example, a MEMS device wherein a sensitive part of the device that forms the membrane is separated from the rest of the chip and supported by springs. The springs decouple the sensitive part from the rest of the chip and absorb the package stress, without transferring it to the sensitive part. In this device, the sensitive part is housed within or faces a second cavity that enables a limited movement of the sensitive part with respect to the rest of the chip.
In practice, the device has two cavities, where a first cavity defines the membrane and a second cavity enables decoupling of the sensitive part of the device from the rest. In the known device, to obtain the two cavities, two semiconductor wafers are used, which are bonded together. If the device is provided with a cap, this is formed in a third wafer, which is also bonded, as discussed hereinafter with reference to FIGS. 1 and 2.
FIG. 1 shows in a simplified way a MEMS sensor 1 formed in a chip 10 of semiconductor material, such as silicon. A cap 11 is fixed to a first face 10A of the chip 10, and a closing region 12 is fixed to a second face 10B of the chip 10 via spacers 26.
The chip 10 comprises a suspended region 13 separated from a peripheral portion 18 of the chip 10 through a trench 14. Elastic elements (also referred to as springs 15) support the sensitive region 13 and connect it mechanically to the peripheral portion 18. The sensitive region 13 houses a buried cavity 16 delimiting a membrane 19. The term “buried cavity” herein refers to an empty area (or an area filled with gas) within a semiconductor material body or chip, which extends at a distance from the two main faces of the body, being separated from these faces by portions of semiconductor and/or dielectric material.
A second cavity 21 extends underneath the sensitive region 13. The sensitive region 13 is provided with a stem 20 (also referred to as Z stopper) extending in the second cavity 21 and limiting oscillation of the sensitive region 13 in the event of impact or stresses that might damage the springs 15.
The cap 11 covers here at the top the entire first face 10A of the chip 10 and protects the latter from the external environment. The cap 11 is fixed via bonding regions 22, for example of metal such as gold, tin, or copper, or polymeric material or a glass material (glass-frit), fixed to the peripheral portion 18 and is thus spaced apart by a gap 23 from the first face 10A due to the thickness of the bonding regions 22. Further, the cap 11 has a through hole 24, which fluidically connects the membrane 19 to the environment that surrounds the chip 10.
The closing region 12 has a protection function during handling of the MEMS sensor 1 (for example, during transport to an assembly system). In general, the closing region 12 is constituted by a second chip housing electronic components, such as an ASIC, but may be constituted by another support, such as a printed-circuit board, or the like. Generally, the closing region 12 has a containment trench 17, to prevent material of the bonding regions 26 from reaching the mobile parts, limiting movement thereof in an undesired way.
By virtue of the second cavity 21, the sensitive region 13 bearing the sensitive part of the MEMS sensor (membrane 19) is free to move within certain limits in a vertical direction (perpendicular to the main extension plane of the chip 10 and thus to the faces 10A, 10B thereof) and is not affected by stress during manufacturing, in particular during packaging, in so far as the sensitive region 13 is mechanically decoupled from the peripheral portion.
The device of FIGS. 1 and 2 is formed by bonding three wafers together. In particular, initially (FIG. 3A) a first wafer 350 of monocrystalline silicon is processed for forming the buried cavities 16 delimiting the membranes 19 at the bottom. Formation of the buried cavities 16 may take place in various ways, for example as described in U.S. Pat. No. 8,173,513. Further, on a first face 350A of the first wafer 350 a gold layer is deposited so as to form first bonding and electrical-connection structures 351. In addition, the first wafer 350 is etched from the front using a silicon etching for laterally defining the trenches 14 and the springs 15.
In parallel, before or after (FIG. 3B), a second wafer 400 of monocrystalline silicon is provided with second bonding and electrical-connection structures 401 having a shape and dimensions that are congruent with those of the first bonding and electrical-connection structures 351. Next, using a resist mask, a deep silicon etch is carried out to form holes 403 and trenches 404. Etching is prolonged so that both the holes 403 and the trenches 404 have a greater depth than the thickness intended for the cap 11 (FIG. 1).
Then (FIG. 3C), the second wafer 400 is flipped over and fixed to the first wafer 350 via a wafer-to-wafer bonding process of a known type, to obtain a composite wafer 500.
Next (FIG. 3D), the first wafer 350 is thinned out from the back, to form the second cavities 21 and the stems 20, and is etched once again from the back, to release the suspended regions 13 and the springs 15. In addition, the second wafer 400 is thinned out until the bottom of the holes 403 and of the trenches 404 is reached.
After bonding a third wafer 410 and dicing the composite wafer 500 of FIG. 3D, the MEMS sensor 1 of FIG. 1 is thus obtained.
Consequently, in the process described, the MEMS device 1 is obtained by bonding three different wafers.
Thus, its thickness is considerable. Further, the process is rather complex in so far as it specifies bonding of three wafers.